Towards Fluid Intake Quantification in Older Adults: An Algorithm for Movement Detection Using Accelerometry and Gyroscope Sensors

Abstract

Dehydration in older adults leads to numerous adverse outcomes and is associated with cognitive diseases and medical treatments affecting thirst sensitivity. Technological solutions, including wearable sensors in devices like smartwatches, play a crucial role in monitoring physical activity and health. The integration of wearable technology, data processing, and advanced analysis offers a non-intrusive, accurate method for quantifying fluid intake in older adults by analyzing wrist and arm movements. This paper presents an algorithm for detecting liquid intake movements in older adults using signal samples collected from a wearable sensor equipped with accelerometry and gyroscope technology. The algorithm focuses on preprocessing the signals and extracting relevant features related to fluid intake in two different experimental conditions: standing and sitting. The proposed algorithm accurately identifies fluid intake events, including instances with uncontrolled movements. Various validation metrics were used to evaluate the algorithm’s performance, including precision, accuracy, sensitivity, and F1-Score. The recorded metrics consistently showed high percentages ranging from 92% to 100%. These results indicate the algorithm’s effectiveness in accurately detecting and classifying fluid intake events in older adults.